[GH-ISSUE #10462] how to run ollama model on CPU instead of GPU even the GPU is available? #32640

Closed
opened 2026-04-22 14:16:52 -05:00 by GiteaMirror · 2 comments
Owner

Originally created by @Bob123Yang on GitHub (Apr 29, 2025).
Original GitHub issue: https://github.com/ollama/ollama/issues/10462

I have 1T system memory for my Ubuntu 22.04 and install Ollama by official method and pull deepseek-r1:671b-q8_0 to the local SSD successfully.

So, how to run ollama model on CPU instead of GPU even the GPU is available?

Originally created by @Bob123Yang on GitHub (Apr 29, 2025). Original GitHub issue: https://github.com/ollama/ollama/issues/10462 I have 1T system memory for my Ubuntu 22.04 and install Ollama by official method and pull deepseek-r1:671b-q8_0 to the local SSD successfully. So, how to run ollama model on CPU instead of GPU even the GPU is available?
GiteaMirror added the question label 2026-04-22 14:16:52 -05:00
Author
Owner

@rick-github commented on GitHub (Apr 29, 2025):

Set num_gpu to 0 as described here.

<!-- gh-comment-id:2837999606 --> @rick-github commented on GitHub (Apr 29, 2025): Set `num_gpu` to 0 as described [here](https://github.com/ollama/ollama/issues/6950#issuecomment-2373663650).
Author
Owner

@pdevine commented on GitHub (Apr 29, 2025):

Going to mark this as answered. Thank you @rick-github !

<!-- gh-comment-id:2839556873 --> @pdevine commented on GitHub (Apr 29, 2025): Going to mark this as answered. Thank you @rick-github !
Sign in to join this conversation.
1 Participants
Notifications
Due Date
No due date set.
Dependencies

No dependencies set.

Reference: github-starred/ollama#32640